The efficacy and safety of an alternating regime with zidovudine and didanosine versus treatment with either drug alone were investigated in a randomized, open, controlled trial, 552 patients with advanced HIV infection, 47% of whom had received prior treatment with zidovudine, were enrolled. The patients were randomly assigned to zidovudine 600 mg/day, didanosine 400 mg/day or 4-week alternations with the 2 drugs in the same dose. The study had a median length of follow-up of 88 weeks. In the overall analyses, time to death (p = 0.48) and time to death or new AIDA event (0.80) were equally distributed between the 3 treatment groups. In the subgroup of patients with a CD4 count < 100 x 10(6)/l the survival was longer in the alternating arm (p < 0.005) primarily because of differences among zidovudine naive patients. The alternating regime was better tolerated than the 2 monotherapies, with a longer time to dose reduction or withdrawal owing to side effects (p < 0.001).